Method for detecting a plurality of instances of an object

Abstract

An improved object recognition method is provided that enables the recognition of many objects in a single image. Multiple instances of an object in an image can now be detected with high accuracy. The method receives a plurality of matches of feature points between a database image and a query image and determines a kernel bandwidth based on statistics of the database image. The kernel bandwidth is used in clustering the matches. The clustered matches are then analyzed to determine the number of instances of the object within each cluster. A recursive geometric fitting can be applied to each cluster to further improve accuracy.

Description

FIELD OF THE DISCLOSURE[0001]The present invention relates generally to object recognition techniques and more particularly to techniques focused on the detection of multiple instances of an object within an image.BACKGROUND[0002]Object recognition systems are typically utilized to find a given object in an image or video sequence. Humans recognize a multitude of objects in images with little effort, despite the fact that the image of the objects may vary somewhat in different viewpoints, in many different sizes / scale or even when the objects are translated or rotated. Objects can even be recognized by humans when the objects are partially obstructed from view. Hence, object recognition systems aim to duplicate the abilities of human vision and understanding of an image.[0003]Object recognition systems are utilized for acquiring, processing, analyzing, and understanding images and, in general, high-dimensional data from the real world in order to produce numerical or symbolic inform...

AI Technical Summary

Benefits of technology

The present invention provides an improved object recognition method that can detect multiple instances of an object within an image with high accuracy. This method involves analyzing feature points in a database image and a query image, and using a kernel bandwidth to cluster the feature points and determine the number of instances of the object in the image. This approach increases the accuracy of object recognition in challenging situations where there are many objects in an image or the objects are partially obstructed. The method can be applied to various object recognition systems and can be used in various fields such as product recognition.

Problems solved by technology

However, a significantly harder use case presents the challenge of recognizing many objects in a single image, with high accuracy.

Products that are displayed and sold on shelves are particularly difficult for current systems to recognize, because these products tend to have similar labeling across a brand, get moved around, and / or become partially blocked by each other.

Current day object recognition algorithms can fail to correctly count the number of product facings.

Due to the lack of some matches of query image feature points to database image feature points (represented by dots 708) the estimated density only returns one local maxima 710, instead of three, thus resulting in false negatives.

The various embodiments have been achieved without the use of template matching or sliding windows which are known to be expensive and inefficient, especially when dealing with large query images.

Images